Which scope below $300 has best. FFT function

[gfmucci]

Which scope below $300 has best FFT function?  How important is bit memory/record length to the quality of this function compared to other specs?

[nctnico]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

[Fungus]

That is the Analog discovery.

Analog Discovery 2. 

[egonotto]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

Hello,

more FFT length give also more dynamic range.

Best regards
egonotto

[nctnico]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

Hello,

more FFT length give also more dynamic range.

Best regards
egonotto

No. You'll always be limited by the number of bits in order to resolve the signal.

[Wolfgang]

Which scope below $300 has best FFT function?  How important is bit memory/record length to the quality of this function compared to other specs?

The budget scopes like Rigol or Siglent are lousy.
Most important is resolution, because that limits your dynamic range.

What I do is sample the tracks into a PC via SCPI and do the FFT there. MUCH better and flexible.

Expensive scopes with a 16bit resolution mode (oversampling plus averaging) create a good FFT, normally (Keysight Infiniium, e.g.).

[rsjsouza]

Which scope below $300 has best FFT function?  How important is bit memory/record length to the quality of this function compared to other specs?

What is the use? If audio, then a soundcard could go a very long way and do a much better job than any real oscilloscope.

[Elasia]

Which scope below $300 has best FFT function?  How important is bit memory/record length to the quality of this function compared to other specs?

The budget scopes like Rigol or Siglent are lousy.
Most important is resolution, because that limits your dynamic range.

What I do is sample the tracks into a PC via SCPI and do the FFT there. MUCH better and flexible.

Expensive scopes with a 16bit resolution mode (oversampling plus averaging) create a good FFT, normally (Keysight Infiniium, e.g.).

I've been doing this method for years although its usually for just spitting out the occasional fancy graph for an exec lol

Something like matlab will run numbers around most scopes.. it just needs the data

[TK]

You can get the GW Instek GDS1054B for $310 at tequipment.  I think it has a very decent FFT.  Can be hacked to over 100MHz.

[egonotto]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

Hello,

more FFT length give also more dynamic range.

Best regards
egonotto

No. You'll always be limited by the number of bits in order to resolve the signal.

Hello,

but increasing the measurement period and the number of samples by a factor of 10
reduces the bin size with that factor and lowers the noise floor by 10 dB

Best regards

[_Wim_]

That is the Analog discovery.

Analog Discovery 2. 

For FFT (and almost every other spec apart from power supplies) the original discovery has the same specs as version 2, but the original one can often be bought on Ebay for less than 100$, which makes it an absolute bargain for the functionality is offers.

[tautech]

Some examples of SDS1202X-E FFT experiments from a few months back:
https://www.eevblog.com/forum/beginners/is-this-oscilloscope-perhaps-the-best-for-the-beginner/msg2985420/#msg2985420

And more earlier in the thread.

[nctnico]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

Hello,

more FFT length give also more dynamic range.

Best regards
egonotto

No. You'll always be limited by the number of bits in order to resolve the signal.

Hello,

but increasing the measurement period and the number of samples by a factor of 10
reduces the bin size with that factor and lowers the noise floor by 10 dB

Best regards

But you still can't see below the noise floor of the front-end and the dynamic range of the ADC. At some point you are looking at the performance of the analog front-end instead of a signal.

[Fungus]

Some examples of SDS1202X-E FFT experiments from a few months back:
https://www.eevblog.com/forum/beginners/is-this-oscilloscope-perhaps-the-best-for-the-beginner/msg2985420/#msg2985420

And more earlier in the thread.

Is that "below $300"?

[khutch004]

Is that "below $300"?

No, it's $379 and less than $360 with a well known discount. The question is, which is more important to you, $60 or good FFT performance? Your answer may legitimately be different than mine but when I considered everything I went with the $499 (before discount) member of that family and FFT performance was one of the big factors in choosing that series.

[Wolfgang]

Which scope below $300 has best FFT function?  How important is bit memory/record length to the quality of this function compared to other specs?

The budget scopes like Rigol or Siglent are lousy.
Most important is resolution, because that limits your dynamic range.

What I do is sample the tracks into a PC via SCPI and do the FFT there. MUCH better and flexible.

Expensive scopes with a 16bit resolution mode (oversampling plus averaging) create a good FFT, normally (Keysight Infiniium, e.g.).

I've been doing this method for years although its usually for just spitting out the occasional fancy graph for an exec lol

Something like matlab will run numbers around most scopes.. it just needs the data

... even a simple Python script can do that if no MATLAB is available

[Fungus]

The question is, which is more important to you, $60 or good FFT performance? Your answer may legitimately be different than mine but when I considered everything I went with the $499 (before discount) member of that family and FFT performance was one of the big factors in choosing that series.

Feel free to upload evidence to the contrary but I'm not convinced that Siglents have good FFT performance. See this thread from the other day:

https://www.eevblog.com/forum/testgear/fnirsi-1013d-100mhz-tablet-oscilloscope/msg3084923/#msg3084923

"Better than a Rigol DS1054Z" isn't much of a yardstick when looking at FFTs.

The GW-Instek is similarly priced to your Siglent but does a much better job. The Micsigs are also better,

[TK]

The question is, which is more important to you, $60 or good FFT performance? Your answer may legitimately be different than mine but when I considered everything I went with the $499 (before discount) member of that family and FFT performance was one of the big factors in choosing that series.

Feel free to upload evidence to the contrary but I'm not convinced that Siglents have good FFT performance. See this thread from the other day:

https://www.eevblog.com/forum/testgear/fnirsi-1013d-100mhz-tablet-oscilloscope/msg3084923/#msg3084923

"Better than a Rigol DS1054Z" isn't much of a yardstick when looking at FFTs.

The GW-Instek is similarly priced to your Siglent but does a much better job. The Micsigs are also better,

The GW-Instek GDS1054B is priced at $310 at tequipment

[Fungus]

The GW-Instek GDS1054B is priced at $310 at tequipment

Anybody on the market for an entry level scope should be grabbing one, it's a complete bargain at that price!

[jjoonathan]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

[tautech]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

Yep, Dave hated it.

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

[nctnico]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

Yep, Dave hated it.

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

Correct. Meanwhile GW Instek has added a spectrum analysis mode as well. But even the normal FFT mode works just fine; it is just that Dave doesn't like GW Instek for some reason. In reality the GW Instek and Siglent in the same price class are on par which isn't a surprise because the basic components are the same.

[TK]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

Yep, Dave hated it.

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

It has not changed a lot in 4 years, has it?

[TK]

There is another video from Dave where he tests a Near Field Probe and if I remember correctly, he gets a very nice match from a GW Instek FFT, compared to the Rigol Spectrum Analyzer that he uses as the model.

[Gandalf_Sr]

Does the GDS1054B have serial decode and triggering?

[Fungus]

Does the GDS1054B have serial decode and triggering?

It does if you copy the relevant firmware modules onto the internal storage. 

It's a modular firmware system and you can copy modules from higher-up Instek models onto lower-down models.

https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/

They also have 4 channels (with per-channel controls) and a proper push button next to the twisty knob - no more pushing twisty knobs to select things.

[Fungus]

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

What's changed since that video?

(...apart from the Rigol DS1054Z getting a firmware update to do FFT from memory instead of "screen")

[TK]

Does the GDS1054B have serial decode and triggering?

It does if you copy the relevant firmware modules onto the internal storage. 

It's a modular firmware system and you can copy modules from higher-up Instek models onto lower-down models.

https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/
They also have 4 channels (with per-channel controls) and a proper push button next to the twisty knob - no more pushing twisty knobs to select things.

It is now included standard (official) with the latest firmware update

[TK]

Does the GDS1054B have serial decode and triggering?

It does if you copy the relevant firmware modules onto the internal storage. 

It's a modular firmware system and you can copy modules from higher-up Instek models onto lower-down models.

https://www.eevblog.com/forum/testgear/possible-gw-instek-gds-1000b-hack/

They also have 4 channels (with per-channel controls) and a proper push button next to the twisty knob - no more pushing twisty knobs to select things.

I am waiting for my unit to arrive to test if the hack still works with the latest firmware.  I had one in the past and I was able to downgrade to a firmware that supported the hack, then you can upgrade from there.  In theory you can hack to 200MHz and I remember measuring signals all the way up to 300MHz or above

[khutch004]

If the GW-Instek is priced similarly to the Siglent then neither are under $300 are they? My point was that if good FFT performance is important to you then you should be willing to spend what it takes to get it. I don't care which brand and model of scope anyone else here buys. But it is better to spend $360 on something that satisfies you than $285 on something you will use for a while and then replace with what you should have bought in the first place. The Siglent I bought met several requirements that I had, not just good FFT performance. But the FFT performance is quite good enough for my needs. If some other brand/model suits you better I am neither surprised nor offended by that. If something else had suited me better I would have bought it instead. This was my first Siglent purchase, I have no emotional attachment to them.

[jjoonathan]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

Yep, Dave hated it.

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

Are siglent's current entry-level scope FFTs still this bad:

https://youtu.be/Cwbwq-AKbPc?t=475

?

[2N3055]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

Yep, Dave hated it.

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

Are siglent's current entry-level scope FFTs still this bad:

https://youtu.be/Cwbwq-AKbPc?t=475

?

No.

[Fungus]

If the GW-Instek is priced similarly to the Siglent then neither are under $300 are they?

What if it's only $310?

https://www.tequipment.net/Instek/GDS-1054B/Digital-Oscilloscopes/?search=true

[nctnico]

And don't forget the Tequipment EEVblog discount. Don't know how much it is exactly but 5% already takes the GDS-1054B below $300.

[TK]

And don't forget the Tequipment EEVblog discount. Don't know how much it is exactly but 5% already takes the GDS-1054B below $300.

Don’t tell me they also have EEVblog discount!  Well, little bit late for me but good news anyway

[rsjsouza]

Ooh, good catch! The GW Instek is in a completely different FFT league from the rigol and siglent.

Yep, Dave hated it.

Really, Fungus links a 4 year old video on FFT shootout and thinks it's relevant to todays offerings ?   

The FFT on the Instek looks alright, but the interface is terrible with having to go back and forth on the buttons to change the center, span and magnitude. I can see why he hated it; I would as well.

[nctnico]

Once you get the hang of it, it is actually is quite productive for doing real measurements. I've used my GW Instek scope's FFT function extensively during the development of a project a couple of years ago.

[gfmucci]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

I don't have an extra computer to devote to the bench and not another $1K for one, so no Discovery 2. 

And yes, use is for audio frequency, distortion testing and diagnostics.   The newish Rigol DS 1202Z ($299; ~$30 less than current Intek previously mentioned) has 24M bit depth which I thought would help with fft.

[TK]

FFT on Rigol is bad. The GDS1054B is less than $300 if you apply the eevblog coupon. The 24Mpts does not help with FFT

[Fungus]

And yes, use is for audio frequency, distortion testing and diagnostics.   The newish Rigol DS 1202Z ($299; ~$30 less than current Intek previously mentioned) has 24M bit depth which I thought would help with fft.

The FFT on that series of Rigols is awful, limited to 64k points and terrible UI. Do not buy one if "use is for audio frequency, distortion testing and diagnostics".

(if that really is the main use for this then get an Analog Discovery instead, it has 14-bit ADC for much lower noise floor and built-in signal generator for bode plots)

[TK]

DS1000Z vs GDS-1054Z FFT comparison video

[Gandalf_Sr]

Yes the GW Instek GDS-1054Z has a 'better' FFT than the Rigol DS1054Z but the UI on the GDS-1054Z is "Ugh" according Dave Jones' video comparing many O'scope brands.



It's like asking which is the better 50" TV to buy when I have only $50 to spend.

My point being that, if you really want to do FFT, you'd be better bumping up your budget to the $600 - $900 range.

If you really want to do spectrum analysis, buy a spectrum analyzer.

[nctnico]

Yes the GW Instek GDS-1054Z has a 'better' FFT than the Rigol DS1054Z but the UI on the GDS-1054Z is "Ugh" according Dave Jones' video comparing many O'scope brands.

The UI on the GW Instek isn't that bad. It is just compact with dual function keys in order to minimize the number of menus to go through. And nowadays GW Instek has a spectrum analysis mode as well.

[Fungus]

Yes the GW Instek GDS-1054Z has a 'better' FFT than the Rigol DS1054Z but the UI on the GDS-1054Z is "Ugh" according Dave Jones' video comparing many O'scope brands.

He also said, "...ultimately, it can do the business".

[Gandalf_Sr]

I'm not bashing GW Instek and I'm certainly not saying that the DS1054Z has a usable FFT; I watched the video and I see why Dave described the UI controls for the FFT function on the GDS-1054Z as "Ugh".  This has fired me up to go into my mancave and test out the FFT FUNctionality of my Rigol MSO5074.

[EDIT] I repeated Dave's FFT test on my MSO5074 and it works like a dream...

[Fungus]

[EDIT] I repeated Dave's FFT test on my MSO5074 and it works like a dream...

Not exactly "below $300" though.

If we're going to be strict about the competition rules then here's a signal with four harmonics being shown on my DSO Quad:



It updates at full screen rate. It also has a built-in signal generator and can do Bode plots. They cost about $160. 


nb. The signal is coming from a digital pin on an Arduino going into an RC filter so S:N ratio is about 48dB.

[Gandalf_Sr]

What's a DSO Quad? The graphics look like an old Pong machine.

[Fungus]

What's a DSO Quad? The graphics look like an old Pong machine.

https://www.ebay.com/sch/i.html?_nkw=dso+quad

[gfmucci]

Yes the GW Instek GDS-1054Z has a 'better' FFT than the Rigol DS1054Z but the UI on the GDS-1054Z is "Ugh" ...

My point being that, if you really want to do FFT, you'd be better bumping up your budget to the $600 - $900 range.

If you really want to do spectrum analysis, buy a spectrum analyzer.

When I receive your $500 check in the mail, I'll consider your recommendation.  Even then I might use the $500 to pay next month's mortgage. 

As for 55" tv's, I got a perfectly fine one for $350 a few months ago.  And yes, if I wanted to be a tv elitist, I know I could spend over a few grand.  I find that some have the same view of scopes.

Consider me in the Fungus camp, or even a few levels below.  I have seen videos by audio techs using their scope to measure >0.1% amplifier harmonic distortion using scopes rated fairly poorly for their fft function.
https://youtu.be/n8hNrDcJMCw

I was looking for something that is considered the top of the heap with regard to fft, maybe a little or somewhat better than what the audio guy was using for under $300.  It looks like the GW Instek fits the bill, with the only downside being a little more knob fiddling being required. 

Thanks for the recommendations.  The GW will be added to my  "maybe" list.

[Mr. Scram]

I've seen a DS1054Z used to pipe data to a PC for FFT a long time ago and I've been surprised this isn't much more common. I see some people talking about Matlab or Python here but there doesn't seem to be any common program or method, despite proper spectrum analyzers still being stupidly expensive. Is anyone willing to share their method or setup?

[nctnico]

IIRC Mechatrommer has written an FFT application for the DS1054Z but the update rate will be very low. The advantage of having FFT on an oscilloscope is that it is -somewhat- realtime with updates less than 1 second apart.

[Fungus]

Thanks for the recommendations.  The GW will be added to my  "maybe" list.

I don't know how long that offer will stand but $300 is not a normal price for the Instek.

Normally they cost a lot more than that.

[Gandalf_Sr]

Knob twiddling is overrated.  Rigol DS1054Z FFT sucks but other than that, it's a great scope.

I struggled with the price point on the TV example, CostCo are selling a 55" 4k TV for $280  shame they don't sell oscilloscopes 

[Antonio90]

That is the Analog discovery.

FFT length is important to get a high frequency resolution. More bits means more dynamic range.

I don't have an extra computer to devote to the bench and not another $1K for one, so no Discovery 2. 

And yes, use is for audio frequency, distortion testing and diagnostics.   The newish Rigol DS 1202Z ($299; ~$30 less than current Intek previously mentioned) has 24M bit depth which I thought would help with fft.

I'm not really in a position to recommend anything, but if you are considering the Analog Discovery, you can probably get a second hand laptop in ebay for ~30-50 bucks working perfectly. My lab computer is a Thinkpad x220 I found for 40€, and second hand computers are usually much more expensive in Europe than in UE. Right now my mother, one of my brothers an a friend of mine are using laptops from 40-60€, it's been a bit more than a year and not a single hardware problem.

[Mr. Scram]

IIRC Mechatrommer has written an FFT application for the DS1054Z but the update rate will be very low. The advantage of having FFT on an oscilloscope is that it is -somewhat- realtime with updates less than 1 second apart.

I've seen a reasonable performing PoC a long time ago. This convinced me a lot should be possible when properly optimized. Unfortunately that software is long gone.

https://hackaday.com/2015/09/22/a-better-spectrum-analyzer-for-your-rigol-scope/

[Wolfgang]

IIRC Mechatrommer has written an FFT application for the DS1054Z but the update rate will be very low. The advantage of having FFT on an oscilloscope is that it is -somewhat- realtime with updates less than 1 second apart.

I've seen a reasonable performing PoC a long time ago. This convinced me a lot should be possible when properly optimized. Unfortunately that software is long gone.

https://hackaday.com/2015/09/22/a-better-spectrum-analyzer-for-your-rigol-scope/

The software is still there, but you need to port it to Python3. It will not do wonders, but as a monitor its not too bad. For a usable FFT you need a high-res scope, otherwise the dynamic range is simply not good enough even for trivial tasks.